Why infrared imagery matters
Many of the most consequential fire–atmosphere interactions occur within and immediately above the flaming zone, where smoke, rapid evolution, and extreme gradients limit the usefulness of conventional visible observations. Thermal infrared sensors penetrate smoke more effectively and provide structured temperature fields suitable for quantitative motion analysis.
What the camera is actually “seeing”
Thermal infrared cameras measure radiance within a wavelength band and convert it to a radiant temperature under assumptions about emissivity and atmospheric transmission. In crown fires, emission from incandescent soot and other hot particles often dominates the signal, forming a textured, dynamically evolving surface suitable for motion tracking.
- Signal source: radiating hot particles and gases
- Advantage: strong spatial gradients through smoke
- Geometry: a projected, warped emitting surface
Pixel size and sampling depend on optics, range, and platform stability. High-frequency imaging (tens of Hz) is essential for resolving sub-second fire–atmosphere coupling without aliasing.
Radiant temperature should be interpreted as a physically meaningful proxy—not a direct thermodynamic temperature of the flame.
End-to-end analysis workflow
- Infrared video acquisition (preferably fixed or stabilized)
- Radiometric normalization or temperature conversion
- Temporal resampling / deinterlacing if required
- Image registration and georeferencing
- Image-flow or gradient-based motion inference
- Derived diagnostics: velocity, spread rate, structure, uncertainty
Deriving winds from infrared imagery
Image-flow methods assume that temperature structure is locally conserved over short time intervals. Spatial gradients of the temperature field, combined with temporal evolution, constrain the two-dimensional velocity field in the image plane.
dT/dt + u (dT/dx) + w (dT/dz) ≈ 0
- Requires sufficient spatial texture
- Robust to partial obscuration by smoke
- Represents projected, not fully 3-D, motion
Fire dynamics revealed by IR analysis
- Counter-rotating convective structures and plume organization
- Near-surface inflow feeding vertical updrafts
- Forward-projecting bursts and “fingers” of flame
- Strong intermittency and sub-second variability
Rate of spread estimation
Sequential airborne and ground-based infrared imagery enables estimation of landscape-scale spread rates when flame-front definition, registration, and temporal sampling uncertainty are explicitly addressed.
Limits and failure modes
- Radiometric ambiguity and emissivity assumptions
- Sensor saturation in high-temperature regions
- 2-D projection of inherently 3-D motion
- Platform motion and video compression artifacts
FROSTFIRE infrared imagery data and analyses
The following MP4 files present data products and derived analyses from sequences of thermal infrared and visible imagery collected during the FROSTFIRE Experimental Burn near Fairbanks, Alaska.
Files with lowercase names primarily represent raw or minimally processed infrared or visible imagery. Files with capitalized names represent derived analysis products (e.g., image-flow, diagnostics, or composited views) computed from the infrared sequences.
- buildie_rgb.mp4
- chopper_rgb.mp4
- escape_2600to12500_6.mp4
- escape_cnt.mp4
- frost_gully2_cnt.mp4
- gully1_cnt.mp4
- gully2_cnt.mp4
- panoram_rgb.mp4
- profil_rgb.mp4
- rollers_rgb.mp4
- IF3_new.mp4 (analysis)
- IF3_P.mp4 (analysis)
- P.mp4 (analysis)
- panoram_test_imgflo.mp4 (analysis)
- PRO.mp4 (analysis)
- profil_IFA_3.mp4 (analysis)
- S.mp4 (analysis)
- SHT.mp4 (analysis)
These materials support the analyses cited below and are provided to allow direct inspection of the infrared structure, motion, and derived diagnostics discussed in the literature.
References
Boroujeni, S.P.H., A. Razi, S. Khoshdel, F. Afghah, J. L. Coen, L. O'Neill, P. Fule, A. Watts, N. T. Kokolakis, K. G. Vamvoudakis, 2024: A comprehensive survey of research towards AI-enabled unmanned aerial systems in pre-, active-, and post-wildfire management. Information Fusion, 108, 102369.
Chen, X., B. Hopkins, H. Wang, L. O'Neill, F. Afghah, A. Razl, P. Fule, J. Coen, E. Rowell, and A. Watts, 2022: Wildland fire detection and monitoring using a drone-collected RGB/IR image dataset. IEEE Access, 10, 121301-121317.
Schag, G. M., D. A. Stow, P. J. Riggan, R. G. Tissell, J. L. Coen, 2021: Examining landscape-scale fuel and terrain controls of wildfire spread rates using repetitive airborne thermal infrared imagery. Fire, 4(1), 6.
Stow, D., P. Riggan, G. Schag, W. Brewer, R. Tissell, J. Coen, and E. Storey, 2019: Assessing uncertainty and demonstrating potential for estimating fire rate of spread at landscape scales based on time sequential airborne thermal infrared imaging. Intl. J. Remote Sensing, 40, 4876-4897.
Coen, J. L., J. Daily, S. Mahalingam, 2010: Application of infrared imagery for understanding wildfire dynamics. Inframation 2010. 15 pp.
Riggan, P. J., L. G. Wolden, R. G. Tissell, J. Coen, 2010: Remote sensing fire and fuels in Southern California. Proc. 3rd Fire Behavior and Fuels Conf.
Coen, J., 2008: Deadly fingers of flame. Southern California Fire Journal, 1, 5-6.
Coen, J. L., S. Mahalingam, and J. W. Daily, 2004: Infrared imagery of crown-fire dynamics during FROSTFIRE. J. Appl. Meteor., 43, 1241-1259.
Radke, L. R., T. L. Clark, J. L. Coen, et al., 2000: The WildFire Experiment (WiFE): Observations with airborne remote sensors. Canadian J. Remote Sensing, 26, 406-417.
Clark, T. L., L. F. Radke, J. L. Coen, and D. Middleton, 1999: Analysis of small-scale convective dynamics in a crown fire using infrared video camera imagery. J. Appl. Meteor., 38, 1401-1420.